Beginners August 2008

beginners@lists.squeakfoundation.org

60 participants
58 discussions

Re: [Newbies] Problems getting started with OpenGL in Squeak
by Ken G. Brown 05 Aug '08

05 Aug '08

At 5:00 AM -0700 8/5/08, beginners-request(a)lists.squeakfoundation.org apparently wrote: >Date: Mon, 4 Aug 2008 14:48:26 -0700 (PDT) >From: Nick Ager <nick.ager(a)gmail.com> >Subject: Re: [Newbies] Problems getting started with OpenGL in Squeak >To: beginners(a)lists.squeakfoundation.org >Message-ID: <18820572.post(a)talk.nabble.com> >Content-Type: text/plain; charset=us-ascii > > > >> Not sure this will help in your case, but just for information, there are >some plugin naming convention i>> ssues, the generated name does not always >match the plugin name. > >>> internal plugin B3DEnginePlugin generated as Squeak3D > >the list produced by executing: "SmalltalkImage current listBuiltinModules" >includes B3DAcceleratorPlugin. Could that be a synonym for "Squeak3D"? Where >is the mapping between the names? > >Looking again at the Croquet image, when showing a 3D window, >"SmalltalkImage current listLoadedModules" includes B3DAcceleratorPlugin and >opengl32.dll > >Although I'm not sure this gets me closer to solving my problem... >-- >View this message in context: http://www.nabble.com/Problems-getting-started-with-OpenGL-in-Squeak-tp1875… >Sent from the Squeak - Beginners mailing list archive at Nabble.com. If you load VMMaker and look in the 'VMMaker' category (package?), you will see InterpreterPlugin class. The plugins are subclasses of this or SmartSyntaxInterpreterPlugin. The #InterpreterPlugin class has the moduleName instance variable. Some of the plugins have class methods that return their name eg. B3DAcceleratorPlugin has moduleName ^'B3DAcceleratorPlugin' Not exactly sure where the moduleName gets set in other plugins. Maybe someone else can tell how this works in more detail. Ken G. Brown

1 0

Re: Tim's Fix for LargeIntger>>AtRandom
by Jerome Peace 05 Aug '08

05 Aug '08

Re: Tim's Fix for LargeIntger>>AtRandom was: [Newbies] Re: What's wrong with this statement? Hi Tim, Glad you are working on a fix. I looked at your first pass. It looks like it would work but maybe eat performance. As Randal mentioned it goes thru Randoms siz and a half times faster than necessary. What happens if you try it with the oddity test I put in the report? Does it work well enough to pass that test? (The data looks ok. Some even some odd. Cool). Feel free to strengthen the test by adding more cases. I would leave the method in Integer as it was but add one in LargePositiveInteger. This would use the heirarchy of classes to make the distinction that is needed. SmallInts and negative large ints would use the current method. The method for large ints would look similar to the way you broke it down. Do a small piece and then the remaining piece. Instead of converting to byte arrays I would just use arithmetic for the recursion. We can choose a large power of two for the radix because each random contains about 2^50 bits of information. LargeInt can probably divide efficiently by powers of two. We don't have to worry about this on the first (uhm, second pass. You did a first pass :-). Once you have a pass that works we could test the next one against it by using a separate generator and noting a beginning seed. (waves hands in the air.) I think I am getting ahead of myself here. outline of nem method: LargePositiveInteger>>atRandom radix := <a truly randomizable power of 2> self > radix ifFalse: [ ^ aGenerator nextInt: self ] . loBits := self \\ radix . hiBits := self // radix . ^ hiBits atRandom * radix + loBits atRandom . (this is a gedanken code to sketch things out. You may have to vary it to get it to work.) I would get something like the above working first. Test it against at least the oddity test. Then maybe play with performace IF you suspected that would be important. If the number is negative the interger method will handle that as it already does. If any of the numbers become Smallints thats covered by what gets done now. For large ints the problem gets broken down to reasonable levels and dealt with. Should work. Hth, Yours is curiosity and service, --Jerome Peace *** >Somebody check me out here, 'cause I think I'm close but something >doesn't seem quite right. > >Attached is a new Integer>atRandom I'm playing with. I break the >integer into a ByteArray and call #atRandom for every byte, using 255 >as the bounding value for all but the first byte. The bounding value >for the first byte is itself minus 1. Concatenated together we get a >byte array representing a random value less than the bounding value. > >I think. Maybe I'm doing this wrong. I feel like the first byte >treatment is wrong but I don't know how to achieve the same >result--without it the random value often exceeds the bounding value. > >I went this route as it avoids doing a bunch of bit shifts I'd >probably get wrong anyway. > >10 timesRepeat: [Transcript show: ((2 raisedTo: 128) atRandom hex); cr] > >gives: > >CFE72E55A8C35387C9932D3235937FE6 >A1760C4F27B161A4D2431C94751A2E65 >D9EC4066D2DBD7A00B17B6FC16D985DC >D3140D928C14D8EA7AEB0CC814A2CA8C >EFB37CECB90EFDCF2F93BC34245118C >1B215016A0812E04A5600FA88373BD7B >7B52EB7F11C0DC822B272B70DCE61890 >1D165F2BAF12E8596D496745A41A4BF4 >E8C62D390B69B26CD9FEAF58BD68C17 >3FBEA83CCB11EEDF6993E9BB60171F35 > >-- Tim > >On Sat, Aug 2, 2008 at 11:00 AM, Cerebus <cerebus2 at gmail.com> wrote: >> Bug submitted. >> >> http://bugs.squeak.org/view.php?id=7142 >> >> I may even attempt a fix. :) >> >> -- Tim >> ***

2 1

Re: [Newbies] Problems getting started with OpenGL in Squeak
by Ken G. Brown 04 Aug '08

04 Aug '08

At 4:53 PM -0700 7/31/08, beginners-request(a)lists.squeakfoundation.org apparently wrote: >Date: Thu, 31 Jul 2008 16:53:42 -0700 (PDT) >From: Nick Ager <nick.ager(a)gmail.com> >Subject: Re: [Newbies] Problems getting started with OpenGL in Squeak >To: beginners(a)lists.squeakfoundation.org >Message-ID: <18765876.post(a)talk.nabble.com> >Content-Type: text/plain; charset=us-ascii > > >Hi Derek, > >thanks for the reply. I've inspected "SmalltalkImage current >listLoadedModules" and "SmalltalkImage current listBuiltinModules" and there >is no sign of "module:'Squeak3D'". I also checked my Croquet installation >and again no sign of "module:'Squeak3D'" - however the Croquet image appears >to have a different installation of OpenGL; there is no sign of Balloon3D*. > >Do I need a "Squeak3D.dll" module is so where would I find it? Google points >me to http://wiki.squeak.org/squeak/2874, which is interesting but that >particularly helpful. > >Are there two separate implementations of OpenGL in Squeak - one in Croquet >and one in "standard" squeak. > >confused > > > >Derek O'Connell-2 wrote: >> >> Try inspecting the following to confirm the module is present: >> >> SmalltalkImage current listLoadedModules >> >> if not then same on: >> >> SmalltalkImage current listBuiltinModules >> >> >> On Wed, Jul 30, 2008 at 2:25 PM, Nick Ager <nick.ager(a)gmail.com> wrote: >>> >>> Hi, >>> >>> I'd like to experiment with OpenGL in Squeak. I found the promisingly >>> titled >>> "Howto - Loading OpenGL in Squeak 3.9" from Jakub Kozisek >>> (http://lists.squeakfoundation.org/pipermail/beginners/2006-November/001451.…) >>> and followed the directions: >>> >>> * install FFI and Balloon3D from SqueakMap. (I used the Universe browser >>> to >>> install FFI version 3.9.1 and Balloon3D version 0.18.2) >>> * install OpenGL-ckl.4.mcz from Monticello repository at >>> http://www.squeaksource.com/OpenGL >>> * apply change set from >>> http://lists.squeakfoundation.org/pipermail/beginners/attachments/20061118/… >>> * ensure squeak.ini contains: B3DXUsesOpenGL = 1 >>> * OpenGLMorph new openInWorld. >>> >>> Unfortunately the instructions resulted in the following error: >>> >>>>>> Error: Error a primitive has failed >>> >>> Examining the stack trace the error was in: >>> >>> B3DPrimitiveRasterizerState>>primInitializeBuffers >>> >>> which reads: >>> B3DPrimitiveRasterizerState>>primInitializeBuffers >>> <primitive: 'b3dInitializeRasterizerState' module:'Squeak3D'> >>> ^self primitiveFailed >>> >>> I ran the shipped FFI tests and all tests passed. >>> I'm running under windows using a VM from SqueakVM-Win32-3.10.6-bin.zip >>> and >>> an image sq3.9.1-7075dev08.07.1 >>> >>> I downloaded Croquet which displayed it's 3D world successfully, but >>> would >>> prefer a minimal OpenGL installation. I tried the Croquet VM with my > >> image, >>> but got the same problem. > >> > >> Any ideas? > >> > >> Nick > >> -- Not sure this will help in your case, but just for information, there are some plugin naming convention issues, the generated name does not always match the plugin name. eg: internal plugin B3DEnginePlugin generated as Squeak3D internal plugin BalloonEnginePlugin generated as B2DPlugin internal plugin BitBltSimulation generated as BitBltPlugin internal plugin DSAPlugin generated as DSAPrims internal plugin DeflatePlugin generated as ZipPlugin internal plugin FFIPlugin generated as SqueakFFIPrims internal plugin KlattSynthesizerPlugin generated as Klatt internal plugin LargeIntegersPlugin generated as LargeIntegers Ken G. Brown

2 1

Colors as class variables...
by Sean Allen 04 Aug '08

04 Aug '08

I'm wondering, the Color handling of different colors like blue etc where it dynamically creates class variables for a number of colors, is that the standard idiom or as the code looks to be kind of old, is it no longer accepted practice. If I needed to do something similar should I be following Color's example or instead creating singletons? What's the best practice in this scenario? Are there smalltalk specific drawbacks to either approach that a beginning wouldn't be readily aware of?

1 0

Re: What's wrong with this statement?
by Cerebus 04 Aug '08

04 Aug '08

Bug submitted. http://bugs.squeak.org/view.php?id=7142 I may even attempt a fix. :) -- Tim

2 2

Re: What's wrong with this statement?
by Jerome Peace 02 Aug '08

02 Aug '08

[Newbies] Re: What's wrong with this statement? HI Tim, Hi Randal Randal, I'm with Tim on this one. Random has no way of knowing how many bits to supply. LargeInteger knows how many it needs and can inquire of random how many it can get. So LargeInteger>>atRandom has the responsibility to get the job done. Asking for enough random numbers for all the bits in question. One random can decide on the high order bit range. The next on the low order range within that etc until you have decide on all the choices. You can employing more than one random generator. That would probably insure better distribution of results. Tim, this is an excelent bug find. Please plant a seed on mantis. We need somewhere to put the patch and the tests for the harvesters. The really important thing is to have a test (sunit) that can be run to validate if an image has the fix or not. Any (or all) of the ones in this thread look like good starts. Yours in curiosity and service, --Jerome Peace *** >Randal L. Schwartz merlyn at stonehenge.com >Thu Jul 31 20:45:44 UTC 2008 > > >>>>>> "Timothy" == Timothy J Miller <tmiller at mitre.org> writes: > Timothy>> It certainly is inconsistent to have an Integer method that doesn't Timothy>> invisibly handle large ints. > Randal>The Integer method is fine. The problem is the lack of bits from the PRNG Randal>from class Random. >

1 0

Problems getting started with OpenGL in Squeak
by Nick Ager 01 Aug '08

01 Aug '08

Hi, I'd like to experiment with OpenGL in Squeak. I found the promisingly titled "Howto - Loading OpenGL in Squeak 3.9" from Jakub Kozisek (http://lists.squeakfoundation.org/pipermail/beginners/2006-November/001451.…) and followed the directions: * install FFI and Balloon3D from SqueakMap. (I used the Universe browser to install FFI version 3.9.1 and Balloon3D version 0.18.2) * install OpenGL-ckl.4.mcz from Monticello repository at http://www.squeaksource.com/OpenGL * apply change set from http://lists.squeakfoundation.org/pipermail/beginners/attachments/20061118/… * ensure squeak.ini contains: B3DXUsesOpenGL = 1 * OpenGLMorph new openInWorld. Unfortunately the instructions resulted in the following error: >>> Error: Error a primitive has failed Examining the stack trace the error was in: B3DPrimitiveRasterizerState>>primInitializeBuffers which reads: B3DPrimitiveRasterizerState>>primInitializeBuffers <primitive: 'b3dInitializeRasterizerState' module:'Squeak3D'> ^self primitiveFailed I ran the shipped FFI tests and all tests passed. I'm running under windows using a VM from SqueakVM-Win32-3.10.6-bin.zip and an image sq3.9.1-7075dev08.07.1 I downloaded Croquet which displayed it's 3D world successfully, but would prefer a minimal OpenGL installation. I tried the Croquet VM with my image, but got the same problem. Any ideas? Nick -- View this message in context: http://www.nabble.com/Problems-getting-started-with-OpenGL-in-Squeak-tp1873… Sent from the Squeak - Beginners mailing list archive at Nabble.com.

2 3

Re: [Newbies] What's wrong with this statement?
by johnps11＠bigpond.com 01 Aug '08

01 Aug '08

The issue is almost certainly in atRandom: Running 30 timesRepeat: [ | a | a:= (2 raisedTo: 57) atRandom. Transcript show: a ; show: ' ' ; show: a even ; cr ]. Gives me 30 falses. The odds of that are less than one in a billion (assuming uniformly distributed integers). This doesn't happen for 2^56 or lower, and does for all 2^n, n>= 57. My guess is that it's either a hardware issue or something about Random. I'm not clever enough to understand PRNGs, so I'll leave it others to work out what the answer is, although I suspect that for more than 56 bits you need a PRNG that uses LargeIntegers and not Floats.

3 3

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

Beginners August 2008